Abstract:
This study investigates the influence of mineralogical composition on the failure mode of
selected rocks under compression in southwestern Nigeria. To achieve the set objectives, the data used for this study were obtained from experimental results generated from granite outcrops at different locations as contained in Saliu, 2008 for the determination of mineralogical composition and mechanical properties of the selected rocks. The laboratory experiments carried out in the Camborne School of Mines’ geotechnical laboratory were conducted on granite samples collected from Ikere, Ewu and Ikole in Ekiti State, Oore and Supare in Ondo State and Awo in Osun State. The petrographical analyses conducted at Camborne School of Mines’ geotechnical laboratory using Scanning Electron Microscope (SEM) and transmitted light
microscope for the determination of mineral contents were adopted for this study. The uniaxial compressive strength tests were carried out on eighteen rock samples using MTS servocontrolled Uniaxial compression testing machine. The strain gauge and acoustic emission monitoring machines attached were used to determine and monitor the stages of failure under compression. The artificial neural network model was used to predict the uniaxial compressive strength (UCS) with seven input parameters. The petrographical analyses revealed that the major minerals present in all the selected granite samples were quartz, feldspar, and biotite while amphibole, titanite, and apatite were found as trace minerals. The average uniaxial compressive st rength values for all selected granite rocks range from 123 MPa to 175 MPa. The crack closure values range from 10 MPa to 40 MPa. The crack initiation values range from 90 MPa to 160 MPa. The results reveal that the higher percentage of feldspar results in a lower crack initiation
stress. The crack damage values range from 90 MPa to 168 MPa while the peak stress values range from 123 MPa to 175 MPa. Results indicate that weathering may significantly affect the crack damage and peak stress and thus the uniaxial compressive strength of the granite. The modulus of elasticity ranges from 32 GPa to 57 GPa. It can be deduced from the study that the pre-existing fractures in Ikere granite quickly closed due to higher feldspar contents than Ikole granite with the highest quartz contents. The result of proposed ANN model shows a very high correlation with the experimentally predicted UCS with the R2 value of 0.873 and the overall error of less than 2%. The economic implication of this study for the mining industry is to ascertain the accurate strength of the rock to be used for the different engineering purposes.
